ÃÍãÏ ÃÈæ ÒäØ
08-03-2012, 04:15 AM
The University of Chicago
Law Review
Volume 79 Winter 2012 Number 1
© 2012 by The University of Chicago
SYMPOSIUM
Making Courts Matter: Politics and the
Implementation of State Supreme Court Decisions
Christopher Berry & Charles Wysong
INTRODUCTION
From Alexander Hamilton’s “least dangerous branch”1 to
Gerald Rosenberg’s “hollow hope,”2 the ability of courts to bring
about important policy change has long been the subject of
controversy. Indeed, even as the success of the new institutionalism
in the 1980s and 1990s has transformed “institutions matter” from a
battle cry into an aphorism, the idea that courts matter still encounters
a line of resistance that has long since fallen away from other branches
of American government. If Rosenberg is extreme in his contention
that “U.S. courts can almost never be effective producers of significant
social reform,”3 his point of view is shared by other scholars who
allow the courts only limited or indirect credit for important policy
reforms or suggest that court decisions merely ratify the prevailing
† Associate Professor, Harris School of Public Policy, University of Chicago.
†† JD 2012, Stanford Law School.
The research presented in this article was previously summarized in Christopher Berry and
Charles Wysong, School-Finance Reform in Red and Blue, 10 Educ Next 63 (Summer 2010),
online at http://educationnext.org/files/EdNext_20103_62.pdf (visited Oct 24, 2011).
1 Federalist 78 (Hamilton), in The Federalist 521, 522 (Wesleyan 1961) (Jacob E. Cooke, ed).
2 See generally Gerald N. Rosenberg, The Hollow Hope: Can Courts Bring About Social
Change? (Chicago 1991).
3 Id at 338.
2 The University of Chicago Law Review [79:1
views of the governing elite.4 While the ranks of such revisionists
appear to be growing, however, they remain opposed and
outnumbered by those maintaining the traditional view of the courts
as, in Justice Hugo Black’s words, “havens of refuge for those who
might otherwise suffer because they are helpless, weak, outnumbered,”
or otherwise vulnerable to majority tyranny.5
Whether the United States courts have, in fact, made important
contributions to social change is ultimately an empirical question,
not a matter of legal theory. Yet, even when examining the same
facts on the ground, the two sides of this debate often reach opposing
conclusions.6 Nowhere is this divergence more apparent than in the
burgeoning literature evaluating the importance of Brown v Board
of Education,7 where one can find scholars arguing that the Court
made virtually no contribution to civil rights or school desegregation,8
others arguing that the case made a pivotal contribution in these
areas and to United States history more generally,9 as well as those in
between assigning the Court an important but more modest role.10
4 See, for example, Michael J. Klarman, Rethinking the Civil Rights and Civil Liberties
Revolutions, 82 Va L Rev 1, 7–18 (1996); Cornell W. Clayton and J. Mitchell Pickerill, The
Politics of Criminal Justice: How the New Right Regime Shaped the Rehnquist Court’s Criminal
Justice Jurisprudence, 94 Georgetown L J 1385, 1411–23 (2006); Michael J. Klarman, How
Great Were the “Great” Marshall Court Decisions?, 87 Va L Rev 1111, 1113–53, 1181–82 (2001);
Robert G. McCloskey, The American Supreme Court 13–25, 224–29 (Chicago 1960); Robert A.
Dahl, Decision-Making in a Democracy: The Supreme Court as a National Policy-Maker, 6 J
Pub L 279, 284–86 (1957).
5 Chambers v Florida, 309 US 227, 241 (1940). For explicit responses to the revisionist
critique, see, for example, Michael McCann, How the Supreme Court Matters in American
Politics: New Institutionalist Perspectives, in Howard Gillman and Cornell Clayton, eds, The
Supreme Court in American Politics 63, 63–76 (Kansas 1999); James F. Spriggs II, The Supreme
Court and Federal Administrative Agencies: A Resource-Based Theory and Analysis of Judicial
Impact, 40 Am J Pol Sci 1122, 1123–32, 1137–43 (1996). See also generally, for example, David
J. Garrow, Hopelessly Hollow History: Revisionist Devaluing of Brown v. Board of Education,
80 Va L Rev 151 (1994).
6 For a collection of essays debating consequences of Supreme Court decisions, see
generally David A. Schultz, ed, Leveraging the Law: Using the Courts to Achieve Social Change
(Peter Lang 1998).
7 347 US 483 (1954).
8 See, for example, Rosenberg, Hollow Hope at 70 (cited in note 2); Gerald N.
Rosenberg, Brown Is Dead! Long Live Brown! The Endless Attempt to Canonize a Case, 80 Va
L Rev 161, 163–67, 171 (1994).
9 See, for example, Paul Finkelman, Book Review, Civil Rights in Historical Context: In
Defense of Brown, 118 Harv L Rev 973, 1017–29 (2005); Morton J. Horwitz, The Warren Court
and the Pursuit of Justice 15 (Hill & Wang 1998); Jack Greenberg, Crusaders in the Courts:
How a Dedicated Band of Lawyers Fought for the Civil Rights Revolution 264–362 (Basic
Books 1994).
10 See, for example, Michael J. Klarman, Brown, Racial Change, and the Civil Rights
Movement, 80 Va L Rev 7, 10–11 (1994); Mark Tushnet, Some Legacies of Brown v. Board of
Education, 90 Va L Rev 1693, 1693–96, 1705–20 (2004).
2012] Making Courts Matter 3
Little progress is likely to be made in this area by continuing to
rehash the same small set of “landmark” Supreme Court decisions.
The focus on singular court decisions has yielded fascinating
historical case studies but little in the way of generalizable empirical
evidence on the role of courts in the policymaking process. This
Article shifts the spotlight to state supreme courts, where the
operation of fifty institutionally comparable policymaking processes
provides an ideal laboratory for studying when and how court rulings
affect outcomes. A movement to reform school funding has spread
through the state court systems, producing dozens of similar cases
over thirty-five years. Because court-based education funding
reforms are both widespread and relatively homogenous in goals and
methods, they provide an excellent setting in which to empirically
explore the ability of courts to enact social change. Comparing
changes in funding between states where the funding system was
ruled unconstitutional with those where it was not, as well as
exploiting variation in the timing of the decisions, allows us to make
a statistical, rather than anecdotal, assessment of the courts’ impact.
Part II of this Article briefly introduces the legal framework and
background of school-finance judgments (SFJs) and surveys the
social science literature on SJF outcomes. Part III explains the data
and statistical strategy employed by this Article. Parts IV through VI
present the basic findings, evidence of a partisan impact on SFJ
outcomes, and the robustness of our results. In sum, our analysis
challenges the idea that courts are ineffectual in initiating policy
change, while providing some of the first quantitative evidence that
implementation is contingent on interactions between courts and
other political institutions.
I. BACKGROUND AND RELATED LITERATURE
The 1973 case of San Antonio Independent School District v
Rodriguez11 closed the door on education-finance reform through the
federal courts and directed plaintiffs to look to their respective state
constitutions for relief.12 Most state constitutions contain education
clauses that require the state to provide an “adequate education” or
a “thorough and efficient” education for all children,13 and plaintiffs
have relied on these provisions in seeking increases in the financial
11 411 US 1 (1973).
12 Id at 58–59.
13 See John C. Eastman, Reinterpreting the Education Clauses in State Constitutions, in
Martin R. West and Paul E. Peterson, eds, School Money Trials: The Legal Pursuit of
Educational Adequacy 55–64 (Brookings 2007).
4 The University of Chicago Law Review [79:1
resources devoted to public schools.14 Although the legal theories and
details vary from case to case, the goal of school-finance lawsuits is
to improve public education by directing additional financial
resources to underperforming and low-income schools.
Beginning with California’s Serrano v Priest15 in 1971, and spurred
by the Rodriguez decision, the constitutionality of school-finance
systems across the states has been under attack for nearly forty years.
By 2005, education reform advocates had filed 139 separate lawsuits in
45 states challenging their education funding systems on constitutional
grounds.16 In 28 of these states, the school-financing system has been
ruled unconstitutional in one or more court challenges.17
While school-finance lawsuits have attracted significant
attention in the legal community18 and a variety of state-specific
studies have been produced,19 nationwide empirical studies of the
effects of school-finance judgments (SFJs) have been relatively few.
Existing nationwide analyses suggest SFJs have reduced within-state
spending inequality by increasing state-financed education spending
for poor districts.20 Though many states do not reduce support for the
richest districts,21 the progressive increase in funding helps narrow
the gap in total spending between the richest and poorest districts.22
One group of authors—Sheila Murray, William Evans, and Robert
14 For an up-to-date state-by-state catalog of school-funding lawsuits, see National
Education Access Network, online at http://www.schoolfunding.info/ (visited Oct 24, 2011).
15 487 P2d 1241 (Cal 1971).
16 This data is based on West and Peterson, School Money Trials at 345–58 (cited in note 13).
17 “Equity” and “Adequacy” School Funding Liability Court Decisions (National Access
Network 2008), online at http://www.schoolfunding.info/litigation/equityandadequacytable.pdf
(visited Oct 24, 2011).
18 See, for example, James E. Ryan and Thomas Saunders, Foreword to Symposium on
School Finance Litigation: Emerging Trends or Dead Ends?, 22 Yale L & Pol Rev 463, 468–80
(2004).
19 See, for example, William S. Koski, The Politics of Judicial Decision-Making in
Educational Policy Reform Litigation, 55 Hastings L J 1077, 1104–1227 (2004).
20 See, for example, Sheila E. Murray, William N. Evans, and Robert M. Schwab,
Education-Finance Reform and the Distribution of Education Resources, 88 Am Econ Rev 789,
806 (1998) (examining data from 1972 to 1992 and finding inequality reduced by 19 to 34 percent).
See also David Card and A. Abigail Payne, School Finance Reform, the Distribution of School
Spending, and the Distribution of Student Test Scores, 83 J Pub Econ 49, 64–67 (2002)
(examining data from 1977 to 1992 and finding that “states in which the supreme court found
the school financing system unconstitutional have altered their funding systems so as to
redistribute aid to lower income districts”).
21 See Murray, Evans, and Schwab, 88 Am Econ Rev at 790 (cited in note 20).
22 See Card and Payne, 83 J Pub Econ at 49 (cited in note 20). David Card and Abigail
Payne also demonstrate that only court decisions in which the school-financing system is found
unconstitutional affect spending, whereas cases in which a state’s system is challenged but
upheld have no measurable effects, suggesting that the decision itself has a causal impact on
funding. Id at 64. They also provide evidence that the equalization of spending resulting from
SFJs leads to a narrowing of test scores across family background groups. Id at 76–79.
2012] Making Courts Matter 5
Schwab—argues that additional state funding for education resulting
from SFJs was financed through higher taxes rather than shifting
resources from other functions,23 although Katherine Baicker and
Nora Gordon find roughly a dollar decrease in other state
intergovernmental transfers to local government, for each dollar
increase in state education spending.24 Card and Payne find that local
governments also respond to increases in state education spending
by reducing their own revenue and spending on education. In other
words, there is evidence that localities “undo” state efforts at
increasing education spending after SFJs. For each dollar of
additional state aid for education, the net increase in total education
funding is fifty to sixty-five cents, after local spending changes are
taken into account.25
Researchers have also begun to explore variation among SFJ
outcomes. Matthew Springer, Keke Liu, and James Guthrie,26 as well
as Christopher Berry,27 attempt to differentiate the effects of equityand
adequacy-based SFJs on spending inequality. Prior to Rose v
Council for Better Education,28 court challenges to school-finance
systems rested on equity grounds.29 Rose, however, ushered in a
series of recent cases challenging state funding systems on adequacy
grounds.30 Although both equity and adequacy cases have usually
been grounded in a demand to provide greater resources for poor
students, adequacy cases acknowledge that resources may be
distributed unequally, to the extent that high-needs students require
23 Murray, Evans, and Schwab, 88 Am Econ Rev at 807 (cited in note 20).
24 See generally Katherine Baicker and Nora Gordon, The Effect of State Education
Finance Reform on Total Local Resources, 90 J Pub Econ 1519 (2006) (examining county-level
data from 1982 to 1997 estimating that SFJs result in a $48 per capita increase in state education
spending, but a $51 per capita reduction in other intergovernmental transfers from the state).
25 Card and Payne, 83 J Pub Econ at 50 (cited in note 20) (noting that increases in state
aid cause local governments to reduce taxes, generating a net increase of only fifty to sixty-five
cents).
26 Matthew G. Springer, Keke Liu, and James W. Guthrie, The Impact of School Finance
Litigation on Resource Distribution: A Comparison of Court-Mandated Equity and Adequacy
Reforms?, 17 Educ Econ 421, 436–41 (2009).
27 Christopher Berry, The Impact of School Finance Judgments on State Fiscal Policy, in
West and Peterson, eds, School Money Trials 213, 222–33 (cited in note 13).
28 790 SW2d 186 (Ky 1989).
29 See Koski, 55 Hastings L J at 1084–87 (cited in note 19).
30 See William S. Koski, The Evolving Role of the Courts in School Reform Twenty Years
after Rose, 98 Ky L J 789, 790–91 (2009). See generally, for example, Davis v South Dakota,
2011 WL 3849864 (SD); Connecticut Coalition for Justice in Education Funding, Inc v Rell,
990 A2d 206 (Conn 2008); Roosevelt Elementary School District No 66 v Bishop, 877 P2d 806
(Ariz 1994); Skeen v Minnesota, 505 NW2d 299 (Minn 1993); McDuffy v Secretary of Executive
Office of Education, 615 NE2d 516 (Mass 1993); Abbot v Burke, 575 A2d 359 (NJ 1990).
6 The University of Chicago Law Review [79:1
greater resources to obtain a satisfactory level of education.31
However, evidence so far does not demonstrate a statistically robust
difference in the fiscal effects of equity and adequacy cases.32
Likewise, Caroline Hoxby explores variation in SFJ outcomes by
abandoning the SFJ dummy-variable approach33 in favor of
computing economically meaningful parameters of school-finance
equalization plans on a state-by-state basis.34 She emphasizes that
SFJs can have quite different effects depending on the price and
income effects they impose, whereas the dummy-variable estimates
capture at best the average effects. In the context of the present
Article, Hoxby’s most important point is that the common SFJ
dummy-variable approach masks significant heterogeneity in the
experiences of different SFJs.35
In the remainder of this Article, we look to politics as one
explanation for the heterogeneous effects of SFJs across states. We
emphasize that a court’s ruling of the existing system as
unconstitutional is only the first step toward reform. Some court
orders provide instructions for how the legislature should fix the
system, but most orders let state politicians provide the remedy by
redesigning the finance system themselves.36 State leaders must then
craft a new finance system that garners the approval of the state
legislature and governor. Even after the court ruling, the reform
must pass through the usual lawmaking process of the state.
Therefore, even states with identical court rulings may end up with
very different reforms depending on how the legislature and the
governor respond to them. We investigate the relationship between
political conditions and court impact by estimating models that allow
the effects of an SFJ to vary depending on the partisan composition
of the state government at the time of the ruling. While this
approach captures only one of the many possible ways that politics
31 For a discussion of the different underlying conceptualizations of resource distribution
in equity and adequacy lawsuits, see Anna Williams Shavers, Rethinking the Equity vs.
Adequacy Debate: Implications for Rural School Finance Reform Litigation, 82 Neb L Rev 133,
146–60 (2003).
32 See Springer, Liu, and Guthrie, 17 Educ Econ at 421 (cited in note 26); Berry, School
Finance Judgments at 222–27 (cited in note 27).
33 Earlier studies included in econometric models a simple on-off switch—a dummy
variable—to indicate whether a court had issued an SFJ in a given state. This approach
captured only the average effect of SFJs and did not distinguish between legal, political,
economic, or other variations in the SFJ decisions or the context surrounding those decisions.
See, for example, Murray, Evans, and Schwab, 88 Econ L Rev at 801–04 (cited in note 20).
34 See Caroline M. Hoxby, All School Finance Equalizations Are Not Created Equal,
116 Q J Econ 1189, 1191–97 (2001).
35 Id at 1192.
36 See, for example, Rose, 790 SW2d at 215.
2012] Making Courts Matter 7
can influence the implementation of a court-ordered reform, our
results suggest that partisan differences are important.
II. DATA AND EMPIRICAL STRATEGY
A. Data
Our analysis examines information from Martin West and Paul
Peterson about SFJs and their outcomes for judgments issued
between 1988 and 2005. We exclude funding lawsuits not related to
general education funding37 and cases that are not the first SFJ in a
given state during our period of study. For the list of cases included
in our analysis, see Appendix Table 1.
Annual district-level financial and demographic information
comes from the Common Core of Data (CCD)38 and the US Census
Bureau Elementary-Secondary School System Finance Survey (F-33)39
when CCD data was not available (1988–1992 and 2005). Our
analysis considers only local school districts and parts of local
supervisory unions with at least one hundred students, as identified
37 We exclude fifteen cases from West and Peterson based on the content of the cases.
For a complete listing of all of the significant school-finance judgments from 1971 to 2005, see
West and Peterson, eds, School Money Trials at 345–58 (cited in note 13). These cases are
Kasayulie v Alaska, 1999 WL 34793400 (Alaska Super) (facilities); Crane Elementary School
District v Arizona, No 1 CA-CV 04-0076 (Ariz Ct App 2006) (capital funding); Hull v Albrecht,
950 P2d 1141 (Ariz 1997) and 960 P2d 634 (Ariz 1998) (capital funding); Roosevelt Elementary
School District Number 66 v Bishop, 877 P2d 806 (Ariz 1994) (capital funding); Haley v
Colorado Department of Education, No 02-CV-5149 (CO, Denver County District Court, 2002)
(special education); Idaho Schools for Equal Educational Opportunity v Idaho, 129 P3d 1199
(Idaho 2005) (facilities); School Administrative District No 1 v Commissioner, Department of
Education, 659 A2d 854 (Me 1995) (upholding method for reducing funding); Durant v
Michigan, 566 NW2d 272 (Mich 1997) (special education); Board of Education of the City of
Millville v New Jersey Department of Education, 872 A2d 1052 (NJ 2005) (preschool); Abbott v Burke,
748 A2d 82 (NJ 2000) (preschool); Zuni Public School District v New Mexico, No CV-98-14-II (NM,
McKinley County District Court, May 26, 1999) (facilities); Tennessee Small School Systems v
McWherter, 894 SW2d 734 (Tenn 1995); Tennessee Small School Systems v McWherter, 91
SW3d 232 (Tenn 2002) (teacher salaries); and Neeley v West Orange-Cove Consolidated
Independent School District, 176 SW3d 746 (Tex 2005) (addressing property taxes, not school
funding). We include two cases that settled with an agreement for reform: Mock v Kansas, No
91-CV-1009 (KS, Shawnee County District Court, Oct 14, 1991), available at Philip C. Kissam,
Constitutional Thought and Public Schools: An Essay on Mock v. State of Kansas, 31
Washburn L J 474, 489–504 (1992); and Minneapolis Branch of the NAACP v Minnesota, No
95-14800 (Minn, Hennepin County Dist Ct 2000), consolidated with Xiong v Minnesota, No 98-
2816 (Minn, Hennepin County Dist Ct 2000).
38 National Center for Education Statistics, Common Core of Data, online at
http://nces.ed.gov/ccd/bat/ (visited Oct 24, 2011) (providing demographic and fiscal information
about public elementary and secondary schools).
39 US Census Bureau, Public School Finance Data, online at http://www.census.gov/govs
/school/ (visited Oct 24, 2011) (providing finance data, including revenues, expenditures, debt,
and assets, for elementary and secondary public school systems).
8 The University of Chicago Law Review [79:1
by the CCD. To address apparent data errors,40 we follow the
approach of Murray, Evans, and Schwab,41 and therefore exclude
districts with total spending per pupil above the ninety-ninth
percentile for each given state-year.42 Additional district
demographic information, such as the proportion of the population
ages five to seventeen and the proportion of school-aged children
living in poverty, comes from the US Census Bureau’s “Model-
Based Small Area Income and Poverty Estimates” for most years.43
Because district poverty information is not available for every year,
the poverty estimates from the closest available survey year are used.
For example, the district poverty estimates for 1996, 1997, and 1998
all use the data from 1997. State-level political control and
demographic information, including population and per capita
income, come from the US Census Bureau’s “Statistical Abstract of
the United States”44 and the National Governors Association
website.45
Local-school-district spending in the United States is financed
by three sources of funding: local revenues raised by the district,
primarily through property taxes; state aid; and federal aid.
Together, state and local funding constitute over 90 percent of school
funding.46 On average, state and local funding each account for
roughly half of this total, although there is substantial variation
across states in the allocation of funding responsibilities between
state and local governments. Most state spending on education
consists of transfers to local districts rather than direct delivery of
40 For example, one district that had 237 students reported over $500,000 in per-pupil
total funding.
41 Murray, Evans, and Schwab, 88 Am Econ Rev at 795 (cited in note 20).
42 We also exclude two districts with implausibly low funding per pupil, as well as Hawaii
and Washington, DC, because each has only one school district.
43 US Census Bureau, Model-Based Small Area Income and Poverty Estimates, online at
http://www.census.gov/did/www/saipe/ (visited Oct 24, 2011). Additionally, for 1989 and 2005,
the district demographic information comes from National Center for Education Statistics,
School District Demographics System, online at http://nces.ed.gov/surveys/sdds/index.aspx
(visited Oct 24, 2011).
44 US Census Bureau, Statistical Abstract of the United States: 2007 253 table 400, online
at http://www.census.gov/prod/2006pubs/07statab/election.pdf (visited Feb 15, 2011); US
Census Bureau, Statistical Abstract of the United States: 1990 259 table 434, online at
http://www2.census.gov/prod2/statcomp/documents/1990-01.pdf (visited Feb 15, 2011).
45 National Governors Association, Past Governors Bios, online at http://www.nga.org/cms
/home/governors/past-governors-bios.html (visited Oct 24, 2011).
46 National Center for Education Statistics, Revenues and Expenditures for Public
Elementary and Secondary Education: School Year 2008–09 (Fiscal Year 2009) 5 (2011), online
at http://nces.ed.gov/pubs2011/2011329.pdf (visited Oct 24, 2011).
2012] Making Courts Matter 9
educational services.47 The federal government has never provided a
significant portion of funding for public education but rather limits
its role to funding specific programs, such as the Free and Reduced
Price Lunch Program, Title I funding for schools that serve students
living in poverty, and various programs supporting special education.
B. Empirical Strategy
In the analyses that follow, we estimate the impact of SFJs on
state, local, and total funding for school districts. While SFJs require
a policy response from the state government, and therefore are
expected to have a direct impact on state funding, they may also
have an indirect effect on local funding—that is, the revenue raised
by local school districts. Indeed, one concern over the efficacy of
SFJs is that local districts may reduce their own taxes in response to
increases in state aid, thereby undermining state-level efforts to
increase school spending. For example, David Card and Abigail
Payne find that a dollar of increased state aid to a local school
district results in only fifty to sixty-five cents in increased spending.48
Therefore, to provide a more comprehensive picture of the effect of
SFJs, we estimate models of both state and local funding. In
addition, models of total funding capture the net effect of SFJs.49
We use a differences-in-differences strategy to identify the effect
of court-ordered education reform on school district funding.
Specifically, we estimate a series of regression models of the basic
form:
= + + ∙ +
∙ + ∙ +
where represents per-pupil funding (in constant 2007 dollars) for
district located in state at time , and and are district- and
year-fixed effects, respectively. With the inclusion of district-fixed
effects, identification in our models comes from within-district
changes in funding following an SFJ. The fixed effects account for
all, observable or unobservable, time-invariant attributes of the
47 A notable exception is Hawaii, which has the nation’s only entirely state-run school
system. See Hawai’i State Department of Education, Introduction, Organization, online at
http://doe.k12.hi.us/about/intro_org.htm (visited Oct 24, 2011).
48 Card and Payne, 83 J Pub Econ 49, 50 (cited in note 20). But see Baicker and Gordon,
90 J Pub Econ at 1532 (cited in note 24).
49 Because federal aid also contributes to local district spending, total funding is not
simply the sum of state and local funding. However, because we do not expect federal funding
to be influenced by a state SFJ—an expectation that is borne out in the data, see Table 2—we
do not estimate additional models for this variable. Nevertheless, estimates of federal funding
are implicit in the difference between results for total funding and the sum of the results for
state and local funding.
10 The University of Chicago Law Review [79:1
district, including the attributes of the state in which the district is
located. Meanwhile, the year-fixed effects purge the estimates of
national trends that affect all districts, such as economic conditions
or changes in federal education policy. Our main variable of interest
is , which is a dummy variable equal to 1 if the state’s schoolfinance
system has been ruled unconstitutional in a court challenge.
Specifically, the dummy variable is set to 1 in the year after the
decision and each subsequent year.50 In most specifications, we also
interact with state partisan control or district poverty, as will be
explained in the analyses that follow. is a vector of time-varying
district characteristics and is a vector of time-varying state
characteristics, both of which will be specified below. Finally, are
the errors, clustered by state, which account for arbitrary correlation
in the residuals among districts within the same state and serial
correlation of repeated observations from the same district.
We include time-varying state-level characteristics that are
expected to influence state funding of education. We control for the
fraction of the state’s population over sixty-five to account for the
possibility that the elderly oppose increases in school spending.51 We
also control for the fraction of the population that is of school age,
which captures aggregate demand for educational services. At the
same time, increasing per-pupil spending is proportionally more
expensive where children constitute a larger fraction of the
population. We control for per capita income in the state, reflecting
“Wagner’s Law” that demand for government services increases with
income52—that is, the elasticity of government spending with respect
to individual income is greater than 1. We also include the square of
state income to allow for the possibility that the demand for
educational spending is not unbounded. Finally, we include two
dummy variables for years of unified Democratic and unified
Republican control, respectively, of the upper chamber, the lower
chamber, and the governorship, which capture partisan differences in
state education funding. The omitted category is divided
government.
We also control for time-varying characteristics of school districts
that are expected to influence either their receipt of state funding or
their propensity to raise their own local revenue. We include the
50 We allow for a one-year lag for the SFJ to take effect because we assume that any
changes in policy made as a result of the decision will be reflected in the next year’s budget at
the earliest.
51 James M. Poterba, Demographic Structure and the Political Economy of Public
Education, 16 J Pol Analysis & Mgmt 48, 49 (1997).
52 Adolf Wagner, Three Extracts on Public Finance, in Richard A. Musgrave and Alan T.
Peacock, eds, Classics in the Theory of Public Finance 1, 7 (Macmillan 1958).
2012] Making Courts Matter 11
number of students in the district to allow for the possibility that
large districts operate differently from small districts, perhaps due to
economies of scale.53 We control for the percentage of the student
population living in poverty, which we expect to make the district
more likely to receive state funding and less able to raise local
funding. The fraction of students with an individualized education
plan (IEP) is an indicator of special needs, which we expect to result
in higher spending. We also include the proportion of the student
population that is African American and the proportion that is
Hispanic. Although we have no reason to believe that these two
variables directly cause changes in education funding, we have access
to relatively few district covariates and include these due to the
possibility that they are correlated with other relevant factors in the
district, such as property values or population growth.54
Descriptive statistics for all the variables used in our analysis are
presented in Table 1.
III. BASELINE RESULTS
We begin by estimating a set of baseline models that assess the
overall impact of SFJs on education funding. Models 1 through 3 of
Table 2 present the within-district estimates for total, state, and local
funding. The point estimates indicate that state funding (Model 2)
increases following an SFJ by about $775 per pupil, while local
funding (Model 3) declines by roughly $400. Total funding (Model 1)
increases by a little over $300.55 This pattern is broadly consistent
with the results shown by Berry, who finds, using state aggregate
data and a longer time series going back to 1971, that SFJs lead to
increasing centralization of education funding to the state level.56
However, none of the SFJ effects estimated in Models 1 through 3
achieve statistical significance. In other words, although on average
SFJs led to increases in total and state funding and reductions in
local funding, there is substantial variation across states and districts
in the impact of SFJs.
53 See Murray, Evans, and Schwab, 88 Am Econ Rev at 803 (cited in note 20). With the
dependent variable measured in per-pupil terms, a negative coefficient on the number of
students indicates economies of scale.
54 For a more general analysis of the role of district racial composition in school-finance
litigation, see generally James E. Ryan, The Influence of Race in School Finance Reform,
98 Mich L Rev 432 (1999).
55 The change in total funding is not exactly equal to the change in state funding plus the
change in local funding because federal funding, which we do not estimate directly, may also
change after an SFJ. As is implicit in this table and those that follow, changes in federal
funding after an SFJ are a negligible part of the total changes in funding.
56 See Berry, School Finance Judgments at 214, 227–31 (cited in note 27).
12 The University of Chicago Law Review [79:1
Models 4 through 6 isolate one important source of variation in
the impact of SFJs by estimating an interaction between the SFJ
dummy and an indicator for district poverty. Because most lawsuits
were aimed, in one way or another, at increasing funding for poor
districts specifically, there is no reason that an SFJ must affect
funding for the average district. Our estimated interaction term
allows for the impact of an SFJ to be different for poor districts
relative to other districts. Specifically, we use an indicator for
whether a district is in the highest quartile, defined relative to its
state and year, in the proportion of students living in poverty.57 All of
the estimated interactions between district poverty and SFJ status
are highly significant statistically, indicating that SFJs led to changes
in funding disproportionately for poor districts. The total effect of
SFJs on the funding of poor districts is the sum of the SFJ main
effect plus the SFJ times poor interaction, which is reported at the
bottom of the table along with the corresponding standard errors.
Taking this sum, one can see that poor districts received $1,260 more
in state funding, $600 less in local funding, and $640 more in total
funding as a result of SFJs, although the estimates for local revenue
are very imprecise. The net increase of $640 in total funding for poor
districts is fairly modest substantively, considering that total perpupil
funding in our sample is a little over $14,000 on average. Put
differently, $640 represents about one-eighth of a standard deviation
in total per-pupil funding. The results for nonpoor districts follow the
same pattern—a shift from local to state revenue resulting in a
modest increase in total funding—but are even smaller in magnitude
and insignificant statistically.
The main effect of the district-poverty dummy variable—which
can be interpreted as the effect of being in the highest poverty
quartile in the absence of an SFJ—is negative and significant for total
revenue (Model 4). Interestingly, the negative differential in total
funding is due to poor districts receiving less state aid (Model 5),
rather than to differences in locally raised revenue (Model 6). Again,
according to the SFJ-poverty interaction, SFJs lead to significant
increases in state funding, as well as total funding, resulting in poor
districts receiving more total revenue after an SFJ, compared both to
nonpoor districts with an SFJ and to poor districts without an SFJ.
57 Continuous measures of poverty yield comparable results, as will be discussed below.
We begin with the dichotomous variable because we have theoretical reasons to expect that
the relationship between the impact of an SFJ and district-level poverty will not be continuous.
That is, we expect an SFJ to result in increases in funding for poor districts, though not
necessarily reductions in funding for other districts.
2012] Making Courts Matter 13
Overall, the results shown in Models 4 through 6 are consistent
with results from prior studies finding that SFJs result in modest
reductions in the inequality of education funding within a state.58 At
the same time, while these prior studies utilize state-aggregate
measures of inequality, we are the first to produce estimates from
district-level data, demonstrating that SFJs generate increases in
state aid to high-poverty districts specifically.
Among the control variables, the point estimates are generally
in the expected direction, although only a few achieve statistical
significance. Districts spend less per student when enrollment
increases, suggesting the existence of economies of scale. Districts
spend more when the proportion of students with an IEP increases.
Given that IEPs generally apply to students with special needs, this
result is unsurprising. Finally, districts receive more funding in the
years when Republicans control the state government.59 This result is
somewhat surprising in light of the popular perception that
Republicans are the party of fiscal conservatism, although it is
possible that increased education funding comes at the expense of
reductions in other expenditures. We will have more to say about
partisan differences in the next Section.
IV. PARTISAN DIFFERENCES
The results presented in the previous Section recapitulate the
major findings from prior studies—namely, that SFJs lead to
transfers of fiscal responsibilities from local to state governments,
increases in the overall level of education funding, and reductions in
funding inequality. In this Section, we investigate whether the impact
of an SFJ varies depending on the partisan control of the state
government at the time of the decision. We focus on patterns of
partisan control at the time of the decision because the sitting
government is the one obliged to craft a policy response to the
court’s order.60 While later governments are also constrained by the
court’s prior rulings, they are not charged with bringing the schoolfunding
system into compliance in the first place. Ultimately,
whether the government in place at the time of the decision leaves
58 See, for example, Berry, School Finance Judgments at 233 (cited in note 27); Card and
Payne, 83 J Pub Econ at 66 (cited in note 20). But see Murray, Evans, and Schwab, 88 Am
Econ Rev at 806 (cited in note 20). High poverty does not necessarily imply low spending, so
the results in Table 2 do not speak directly to the issue of spending inequality.
59 See Table 2.
60 Of course, there are instances in which the sitting legislature and governor fail to agree
on a response to the court’s order before their term is up, thereby leaving the problem for their
successors to tackle. We discuss these issues below.
14 The University of Chicago Law Review [79:1
an enduring mark on the system is an empirical question, and later
we will discuss the robustness of our results as compared to
alternative ways of modeling partisan influence.
We begin by creating three dummy variables representing,
respectively, unified Democratic control of the state legislature and
governorship at the time of the court decision, unified Republican
control, and divided government. For example, the first dummy
variable will equal 1 for years after an SFJ if Democrats controlled
both houses of the legislature and the governorship in the year of the
decision, 0 otherwise. This approach allows us to test whether the
within-district difference in funding before versus after an SFJ is
different depending on the partisan composition of the state
government at the time of the decision. The omitted category is
again no SFJ. Thus, identification in these models comes both from
within-district variation (to estimate the change in spending after an
SFJ) and from between-state variation (to estimate the differential
effects of the three configurations of partisan control). We continue
to include the full set of state and district control variables used in
Table 2, although to conserve space we do not report their results.
Appendix Table 1 lists the cases used in our analysis and the
configuration of partisan control of the state government at the time
of the decision. From the outset, we caution that there are only three
SFJs issued during periods of unified Republican government.61
There are seven SFJs handed down to unified Democratic
governments,62 and eleven SFJs delivered to divided governments.63
While we have nearly 200,000 district-level observations in our data
set, our ability to test for partisan differences in outcomes is limited
by the number of cases decided under each regime.
Models 1 through 3 of Table 3 indicate that SFJs handed down
to Democratic state governments result in the greatest changes in
education funding. Specifically, when Democrats are in control at the
time of an SFJ, state revenue subsequently increases by $1,750 per
pupil, while local revenue declines by $630, resulting in an increase in
total funding of $1,150. Meanwhile, SFJs implemented under divided
government result in roughly offsetting increases in state funding and
decreases in local funding, with negligible average effects on total
funding. Finally, when a Republican-controlled government
responds to an SFJ, the result is modest increases in both state and
61 These SFJs were issued for New Hampshire, Ohio, and Wyoming.
62 These SFJs were handed down to Alabama, Kentucky, Maryland, Missouri, Tennessee,
Vermont, and West Virginia.
63 These SFJs were delivered to Connecticut, Idaho, Kansas, Massachusetts, Minnesota,
Montana, North Carolina, New Jersey, New York, South Carolina, and Texas.
2012] Making Courts Matter 15
local funding, summing to a total increase in funding of $365 per
student. Most of the coefficients are imprecisely estimated, however.
Only the effect of Democratic reform on total funding is significant
at the conventional 5 percent level, while the Republican effect on
total funding is significant at 10 percent.
The hypothesis tests implicit in Table 3 indicate whether each
type of partisan reform results in a difference in spending compared
with not having had an SFJ. It is also important to test whether the
different configurations of partisan control have different effects
from each other. These hypothesis tests are reported in Table 3A.
The evidence shows that Democratic-controlled reforms do not have
the same total effect as a reform implemented under divided
government. However, all of the other tests fail to reject the
hypothesis that reforms issued under different partisan
configurations have equal effects.
The imprecision of the estimated effects in Models 1 through 3
again suggests that there is substantial heterogeneity across districts
in the effects of SFJs, even under the same regime of partisan
control. In Models 4 through 6, we allow the effects of SFJs to differ
for poor and nonpoor districts within each regime type. That is, we
have a quadruple-difference model in which we estimate the
difference in the change in spending after an SFJ between poor and
nonpoor districts across three different regimes of partisan control.
Using this strategy, it becomes apparent that the reforms
implemented by Democratic state governments have significantly
different effects from those implemented by divided or Republican
governments.
First, note that the Democratic-reform-times-poor-districtinteraction
term is substantively small and statistically insignificant
for total revenue, indicating that poor districts do not fare differently
than other districts under a Democratically controlled SFJ. The main
effect of Democratic reform remains positive and significant. In
other words, all districts—poor and nonpoor alike—see increases in
funding following an SFJ implemented by a Democratic state
government. According to Models 5 and 6, poor districts do experience
a greater shift from local to state funding than nonpoor districts
under a Democratic SFJ, but the decrement to local revenue
essentially offsets the increment to state revenue, hence the null
result for total revenue.
By contrast, SFJs implemented by divided or Republican state
governments generate significant increases in funding for poor
districts, relative to other districts. That is, the main effects for split
and Republican SFJs are insignificant, while the interactions with the
16 The University of Chicago Law Review [79:1
poor-district indicator are positive and significant. Under both types
of regimes, poor districts see reductions in local funding that are
more than offset by increases in state funding, resulting in increases
in total funding. Although the point estimate for the SFJ times poordistrict
interaction is larger for Republican governments ($622) than
for divided governments ($498), we cannot reject the hypothesis that
they are equal, as shown in Table 3A.
In sum, an SFJ under each type of partisan regime results in a
change to the school-funding system after the ruling, and
Democratically implemented SFJs have significantly different effects
from SFJs implemented under Republican or divided state
governments. Specifically, Democratic SFJs result in increases in
total funding for both poor and nonpoor districts, whereas a
Republican or divided government SFJ results in increases in
funding for poor districts disproportionately. All three types of SFJs
lead to a shift in funding responsibility from local to state
governments, though to differing degrees.
Thus far in our analysis, we have lumped together all nonpoor
districts within a state into a single category. To produce a more
comprehensive picture of the partisan effects of SFJs, we next extend
our analysis to include interactions between SFJ status, partisan
regime type at the time of the decision, and indicators for four
quartiles of the within-state distribution of district-level poverty. In
other words, we divide each state’s districts into four quartiles based
on the proportion of students in poverty and then allow each of the
three different types of SFJs—Democratic, Republican, and divided
state government—to have differential effects for each quartile. This
allows us to assess, in a nonparametric fashion, the effect of an SFJ
across the spectrum of district poverty. As before, the inclusion of
district- and year-fixed effects means that our estimates of the effect
of an SFJ come from within-district changes in spending after the
decision. We simply compare these within-district changes across
different combinations of partisan control and district poverty.
The results of this analysis are presented in Table 4. The omitted
category of district poverty is the second-poorest quartile, so the
coefficients for the remaining quartiles should be interpreted as
differences with respect to this omitted category. However, these
complex interactions are more easily presented graphically, which we
do in Figure 1. Each bar in the graph represents the total effect of an
SFJ—that is, the within-district change in spending after the
decision—for each category of partisan control and district poverty,
computed from the coefficients shown in Table 4. Meanwhile,
2012] Making Courts Matter 17
Table 4A presents associated tests for differences in the effects of
SFJs across various partisan and poverty groups.
The more detailed results depicted in Figure 1 indicate that SFJs
crafted by Democratic governments result in total funding increases
across the board. Districts in every poverty quartile see a shift from
local to state funding after an SFJ, with a net increase in funding
ranging from roughly $1,000 to $1,300 per pupil. While a few of the
differences between quartiles are significant statistically, they are
substantively small relative to the overall level of the funding
increases. Indeed, if anything, the results indicate that the most
affluent districts fare better than the poorest districts, in terms of
total funding, after a Democratic SFJ, although this difference is not
statistically significant.
SFJ reforms implemented under divided government produce
decidedly different results. There is a shift from local to state
funding, but the change in state funding is monotonically increasing
district poverty. The poorest quartile of districts receives significantly
greater increases in state funding than any other quartile. Moreover,
the poorest quartile is the only one to see positive net changes in
total funding. Overall, divided government SFJs appear to represent
a more or less straightforward redistribution of funding toward the
poorest districts. The net effects on state education funding appear
to be neutral, as further evidenced by Model 1 of Table 3, showing
little change in total education funding after an SFJ under divided
government. That said, the net increase in total funding for poor
districts is substantively small, amounting to roughly $350 per
student, or only 7 percent of a standard deviation in total funding.
Republican-controlled SFJs present yet a third distinct pattern
of funding reforms. Under Republican reforms, there is a shift from
local to state funding only for the poorest districts. Districts in the
most affluent quartile face cuts in state funding, but they are able to
more than compensate for these reductions by increasing their own
local revenue. In other words, Republican SFJs involve increasing
centralization of funding for the poorest districts and increasing
decentralization of funding for the richest districts. The middle two
quartiles are essentially unaffected. On net, both the poorest and the
richest districts see increases in total funding, the former courtesy of
state aid and the latter financed from their own tax base. Indeed, the
richest districts in Republican states are the only quartile under any
partisan regime to experience an increase in local funding following
an SFJ.
Which partisan arrangement leads to the best results for poor
districts after an SFJ? Clearly, SFJs implemented by Democrats
18 The University of Chicago Law Review [79:1
produce the largest net increases in funding. However, by delivering
roughly equivalent funding increases to districts at all income levels,
Democratic SFJs will do little to reduce the overall level of funding
inequality in their states. On the other hand, SFJs implemented by
divided or Republican governments deliver concentrated benefits to
poor districts, thereby reducing, albeit moderately, the funding
differential between poor and more affluent districts. However,
especially under divided government SFJs, the actual flow of new
dollars into poor districts is more meager than when Democrats are
in control. Assessing which sort of reform produces the greatest
benefits would require a richer theory and empirical analysis of the
connection between student outcomes and the level and inequality of
school funding.
We do not offer a substantive political explanation for the
differential partisan responses to state SFJs. One simple conjecture
would be that political actors use the court rulings as a pretense to
enact policy changes that they were otherwise predisposed to make
anyway. Yet, it is not clear why such a pretense would be necessary
in a state dominated by one party. For example, when Democrats
control the legislature and governorship, they presumably will set the
policy they want absent court intervention, so it is not obvious why
an SFJ should elicit a greater response from Democrats than from
Republicans. It is possible that supermajoritarian voting requirements
within the legislature prevent the dominant party from attaining its
ideal policies even under unified government, and that court rulings
help surmount such hurdles. However, a comprehensive account of
court influence over state policymaking is beyond the scope of this
paper and so we leave the theoretical underpinnings of our results as
fodder for future research.
V. ROBUSTNESS
In additional analyses not reported, we have explored the
robustness of our results to various changes in model specification
and variable definitions. First, we examined sensitivity to our
treatment of partisan control. In the models reported above, we
coded each SFJ according to the partisan alignment in the year of the
decision. As an alternative, we also reran the models using
contemporaneous measures of partisan control—that is, interacting
the SFJ dummy with indicators of partisan control in the current year
for all years after the decision. The signs of the coefficients were
unchanged, but they became much less precisely estimated when
using the contemporaneous measures of partisanship. This leads us
to conclude that state party control at the time of the decision
2012] Making Courts Matter 19
captures an important determinant of the political response to the
court decision.
As noted in our discussion of the school-finance litigation
movement, many states have had more than one court ruling on
education finance.64 The rulings that determine our coding of the SFJ
dummy variable are listed in Appendix Table 1. However, some
states had rulings in place prior to the beginning of our study period,
and some states had additional rulings that came later. We have
explored various sensitivity analyses as to the coding of the SFJ
dummy. For example, we recoded the dummy variable to reflect
cases decided before the beginning of our study period and reran the
models. We also tried including indicator variables for the existence
of reform decisions that came before or after the cases used in our
analysis. Our results did not change importantly, although we did
observe that the first reform case tends to have the largest effect on
fiscal outcomes.
A lingering concern with our results may be that the filing of
lawsuits is endogenously related to political control of the state
government. First, we note that because nearly all states—forty-five
of fifty—were subject to at least one education-finance lawsuit, the
central issue is not whether a state would face a suit but when. One
possibility is that education-reform advocates may time the filing of
their lawsuits to take advantage of what they view as particularly
favorable political conditions. Another possibility is that advocates
might resort to litigation only when the legislative and political
process fails to provide reform. Therefore, endogenous timing by
litigants might potentially bias our results in either direction.
We believe that the possibility of endogenous lawsuit timing visà-
vis political regime is not a major concern for three reasons: the
amount of time between lawsuit filing and the court decision is
unpredictable, political regimes often change between the lawsuit
filing and decision, and lawsuits do not appear to be precipitated by
changes in political regime. Among the twenty-three cases that
contribute to our estimate of SFJ reform, the length of time from the
initial filing through the final appellate court decision ranged from
less than a year to nine years. On average, the process took four
years. Due to the length of time the suits take and the variability of
the speed of the adjudication process, advocates could not effectively
time their lawsuits to specific political regimes. In almost half of the
cases—eleven out of twenty-three—the political regime changed
between the time of filing and time of decision. When political
64 See note 16 and accompanying text.
20 The University of Chicago Law Review [79:1
control changes more rapidly than lawsuits can be adjudicated,
reform advocates cannot time their lawsuits to be decided under a
specific regime. Further, SFJ lawsuits do not appear to be motivated
by a change in regime preceding case filing. On average, the political
regime in the state was stable for six years prior to the filing of a
case. In only three cases did the regime change in the year of the
lawsuit filing, and for each of those three cases, the political regime
changed again before the lawsuit was decided. Thus, we do not see
evidence that endogenous timing of lawsuits is likely to be an
important source of bias for our results.
CONCLUSION
The state experience with school-finance litigation provides
evidence that courts have played a significant role in at least this one
realm of social reform. We find statistically significant differences in
education funding following court judgments overturning a state’s
school-financing system. In particular, Democratic, Republican, and
divided state governments produce significantly different education
spending patterns following an SFJ, once we account for the relative
poverty of the school districts. Further, SFJ reforms overall had a
significantly different impact on low-spending school districts than
on high-spending school districts. The effects of a court decision are
contingent on the pattern of partisan control in place at the time.
Thus, our analysis challenges the idea that courts are ineffectual in
initiating policy change while providing some of the first quantitative
evidence that implementation is contingent on interactions between
courts and other political institutions.
2012] Making Courts Matter 21
FIGURE 1. EFFECT OF SFJ ON SCHOOL FUNDING BY PARTY
CONTROL AND POVERTY QUARTILE
Notes: This is a graphical depiction of the results shown in Table 4. Bars represent the sum of the
coefficients for the partisan SFJ main effect plus its interaction with district poverty. The first bar in each
section indicates the highest poverty quartile, while the fourth bar indicates the lowest poverty quartile.
Law Review
Volume 79 Winter 2012 Number 1
© 2012 by The University of Chicago
SYMPOSIUM
Making Courts Matter: Politics and the
Implementation of State Supreme Court Decisions
Christopher Berry & Charles Wysong
INTRODUCTION
From Alexander Hamilton’s “least dangerous branch”1 to
Gerald Rosenberg’s “hollow hope,”2 the ability of courts to bring
about important policy change has long been the subject of
controversy. Indeed, even as the success of the new institutionalism
in the 1980s and 1990s has transformed “institutions matter” from a
battle cry into an aphorism, the idea that courts matter still encounters
a line of resistance that has long since fallen away from other branches
of American government. If Rosenberg is extreme in his contention
that “U.S. courts can almost never be effective producers of significant
social reform,”3 his point of view is shared by other scholars who
allow the courts only limited or indirect credit for important policy
reforms or suggest that court decisions merely ratify the prevailing
† Associate Professor, Harris School of Public Policy, University of Chicago.
†† JD 2012, Stanford Law School.
The research presented in this article was previously summarized in Christopher Berry and
Charles Wysong, School-Finance Reform in Red and Blue, 10 Educ Next 63 (Summer 2010),
online at http://educationnext.org/files/EdNext_20103_62.pdf (visited Oct 24, 2011).
1 Federalist 78 (Hamilton), in The Federalist 521, 522 (Wesleyan 1961) (Jacob E. Cooke, ed).
2 See generally Gerald N. Rosenberg, The Hollow Hope: Can Courts Bring About Social
Change? (Chicago 1991).
3 Id at 338.
2 The University of Chicago Law Review [79:1
views of the governing elite.4 While the ranks of such revisionists
appear to be growing, however, they remain opposed and
outnumbered by those maintaining the traditional view of the courts
as, in Justice Hugo Black’s words, “havens of refuge for those who
might otherwise suffer because they are helpless, weak, outnumbered,”
or otherwise vulnerable to majority tyranny.5
Whether the United States courts have, in fact, made important
contributions to social change is ultimately an empirical question,
not a matter of legal theory. Yet, even when examining the same
facts on the ground, the two sides of this debate often reach opposing
conclusions.6 Nowhere is this divergence more apparent than in the
burgeoning literature evaluating the importance of Brown v Board
of Education,7 where one can find scholars arguing that the Court
made virtually no contribution to civil rights or school desegregation,8
others arguing that the case made a pivotal contribution in these
areas and to United States history more generally,9 as well as those in
between assigning the Court an important but more modest role.10
4 See, for example, Michael J. Klarman, Rethinking the Civil Rights and Civil Liberties
Revolutions, 82 Va L Rev 1, 7–18 (1996); Cornell W. Clayton and J. Mitchell Pickerill, The
Politics of Criminal Justice: How the New Right Regime Shaped the Rehnquist Court’s Criminal
Justice Jurisprudence, 94 Georgetown L J 1385, 1411–23 (2006); Michael J. Klarman, How
Great Were the “Great” Marshall Court Decisions?, 87 Va L Rev 1111, 1113–53, 1181–82 (2001);
Robert G. McCloskey, The American Supreme Court 13–25, 224–29 (Chicago 1960); Robert A.
Dahl, Decision-Making in a Democracy: The Supreme Court as a National Policy-Maker, 6 J
Pub L 279, 284–86 (1957).
5 Chambers v Florida, 309 US 227, 241 (1940). For explicit responses to the revisionist
critique, see, for example, Michael McCann, How the Supreme Court Matters in American
Politics: New Institutionalist Perspectives, in Howard Gillman and Cornell Clayton, eds, The
Supreme Court in American Politics 63, 63–76 (Kansas 1999); James F. Spriggs II, The Supreme
Court and Federal Administrative Agencies: A Resource-Based Theory and Analysis of Judicial
Impact, 40 Am J Pol Sci 1122, 1123–32, 1137–43 (1996). See also generally, for example, David
J. Garrow, Hopelessly Hollow History: Revisionist Devaluing of Brown v. Board of Education,
80 Va L Rev 151 (1994).
6 For a collection of essays debating consequences of Supreme Court decisions, see
generally David A. Schultz, ed, Leveraging the Law: Using the Courts to Achieve Social Change
(Peter Lang 1998).
7 347 US 483 (1954).
8 See, for example, Rosenberg, Hollow Hope at 70 (cited in note 2); Gerald N.
Rosenberg, Brown Is Dead! Long Live Brown! The Endless Attempt to Canonize a Case, 80 Va
L Rev 161, 163–67, 171 (1994).
9 See, for example, Paul Finkelman, Book Review, Civil Rights in Historical Context: In
Defense of Brown, 118 Harv L Rev 973, 1017–29 (2005); Morton J. Horwitz, The Warren Court
and the Pursuit of Justice 15 (Hill & Wang 1998); Jack Greenberg, Crusaders in the Courts:
How a Dedicated Band of Lawyers Fought for the Civil Rights Revolution 264–362 (Basic
Books 1994).
10 See, for example, Michael J. Klarman, Brown, Racial Change, and the Civil Rights
Movement, 80 Va L Rev 7, 10–11 (1994); Mark Tushnet, Some Legacies of Brown v. Board of
Education, 90 Va L Rev 1693, 1693–96, 1705–20 (2004).
2012] Making Courts Matter 3
Little progress is likely to be made in this area by continuing to
rehash the same small set of “landmark” Supreme Court decisions.
The focus on singular court decisions has yielded fascinating
historical case studies but little in the way of generalizable empirical
evidence on the role of courts in the policymaking process. This
Article shifts the spotlight to state supreme courts, where the
operation of fifty institutionally comparable policymaking processes
provides an ideal laboratory for studying when and how court rulings
affect outcomes. A movement to reform school funding has spread
through the state court systems, producing dozens of similar cases
over thirty-five years. Because court-based education funding
reforms are both widespread and relatively homogenous in goals and
methods, they provide an excellent setting in which to empirically
explore the ability of courts to enact social change. Comparing
changes in funding between states where the funding system was
ruled unconstitutional with those where it was not, as well as
exploiting variation in the timing of the decisions, allows us to make
a statistical, rather than anecdotal, assessment of the courts’ impact.
Part II of this Article briefly introduces the legal framework and
background of school-finance judgments (SFJs) and surveys the
social science literature on SJF outcomes. Part III explains the data
and statistical strategy employed by this Article. Parts IV through VI
present the basic findings, evidence of a partisan impact on SFJ
outcomes, and the robustness of our results. In sum, our analysis
challenges the idea that courts are ineffectual in initiating policy
change, while providing some of the first quantitative evidence that
implementation is contingent on interactions between courts and
other political institutions.
I. BACKGROUND AND RELATED LITERATURE
The 1973 case of San Antonio Independent School District v
Rodriguez11 closed the door on education-finance reform through the
federal courts and directed plaintiffs to look to their respective state
constitutions for relief.12 Most state constitutions contain education
clauses that require the state to provide an “adequate education” or
a “thorough and efficient” education for all children,13 and plaintiffs
have relied on these provisions in seeking increases in the financial
11 411 US 1 (1973).
12 Id at 58–59.
13 See John C. Eastman, Reinterpreting the Education Clauses in State Constitutions, in
Martin R. West and Paul E. Peterson, eds, School Money Trials: The Legal Pursuit of
Educational Adequacy 55–64 (Brookings 2007).
4 The University of Chicago Law Review [79:1
resources devoted to public schools.14 Although the legal theories and
details vary from case to case, the goal of school-finance lawsuits is
to improve public education by directing additional financial
resources to underperforming and low-income schools.
Beginning with California’s Serrano v Priest15 in 1971, and spurred
by the Rodriguez decision, the constitutionality of school-finance
systems across the states has been under attack for nearly forty years.
By 2005, education reform advocates had filed 139 separate lawsuits in
45 states challenging their education funding systems on constitutional
grounds.16 In 28 of these states, the school-financing system has been
ruled unconstitutional in one or more court challenges.17
While school-finance lawsuits have attracted significant
attention in the legal community18 and a variety of state-specific
studies have been produced,19 nationwide empirical studies of the
effects of school-finance judgments (SFJs) have been relatively few.
Existing nationwide analyses suggest SFJs have reduced within-state
spending inequality by increasing state-financed education spending
for poor districts.20 Though many states do not reduce support for the
richest districts,21 the progressive increase in funding helps narrow
the gap in total spending between the richest and poorest districts.22
One group of authors—Sheila Murray, William Evans, and Robert
14 For an up-to-date state-by-state catalog of school-funding lawsuits, see National
Education Access Network, online at http://www.schoolfunding.info/ (visited Oct 24, 2011).
15 487 P2d 1241 (Cal 1971).
16 This data is based on West and Peterson, School Money Trials at 345–58 (cited in note 13).
17 “Equity” and “Adequacy” School Funding Liability Court Decisions (National Access
Network 2008), online at http://www.schoolfunding.info/litigation/equityandadequacytable.pdf
(visited Oct 24, 2011).
18 See, for example, James E. Ryan and Thomas Saunders, Foreword to Symposium on
School Finance Litigation: Emerging Trends or Dead Ends?, 22 Yale L & Pol Rev 463, 468–80
(2004).
19 See, for example, William S. Koski, The Politics of Judicial Decision-Making in
Educational Policy Reform Litigation, 55 Hastings L J 1077, 1104–1227 (2004).
20 See, for example, Sheila E. Murray, William N. Evans, and Robert M. Schwab,
Education-Finance Reform and the Distribution of Education Resources, 88 Am Econ Rev 789,
806 (1998) (examining data from 1972 to 1992 and finding inequality reduced by 19 to 34 percent).
See also David Card and A. Abigail Payne, School Finance Reform, the Distribution of School
Spending, and the Distribution of Student Test Scores, 83 J Pub Econ 49, 64–67 (2002)
(examining data from 1977 to 1992 and finding that “states in which the supreme court found
the school financing system unconstitutional have altered their funding systems so as to
redistribute aid to lower income districts”).
21 See Murray, Evans, and Schwab, 88 Am Econ Rev at 790 (cited in note 20).
22 See Card and Payne, 83 J Pub Econ at 49 (cited in note 20). David Card and Abigail
Payne also demonstrate that only court decisions in which the school-financing system is found
unconstitutional affect spending, whereas cases in which a state’s system is challenged but
upheld have no measurable effects, suggesting that the decision itself has a causal impact on
funding. Id at 64. They also provide evidence that the equalization of spending resulting from
SFJs leads to a narrowing of test scores across family background groups. Id at 76–79.
2012] Making Courts Matter 5
Schwab—argues that additional state funding for education resulting
from SFJs was financed through higher taxes rather than shifting
resources from other functions,23 although Katherine Baicker and
Nora Gordon find roughly a dollar decrease in other state
intergovernmental transfers to local government, for each dollar
increase in state education spending.24 Card and Payne find that local
governments also respond to increases in state education spending
by reducing their own revenue and spending on education. In other
words, there is evidence that localities “undo” state efforts at
increasing education spending after SFJs. For each dollar of
additional state aid for education, the net increase in total education
funding is fifty to sixty-five cents, after local spending changes are
taken into account.25
Researchers have also begun to explore variation among SFJ
outcomes. Matthew Springer, Keke Liu, and James Guthrie,26 as well
as Christopher Berry,27 attempt to differentiate the effects of equityand
adequacy-based SFJs on spending inequality. Prior to Rose v
Council for Better Education,28 court challenges to school-finance
systems rested on equity grounds.29 Rose, however, ushered in a
series of recent cases challenging state funding systems on adequacy
grounds.30 Although both equity and adequacy cases have usually
been grounded in a demand to provide greater resources for poor
students, adequacy cases acknowledge that resources may be
distributed unequally, to the extent that high-needs students require
23 Murray, Evans, and Schwab, 88 Am Econ Rev at 807 (cited in note 20).
24 See generally Katherine Baicker and Nora Gordon, The Effect of State Education
Finance Reform on Total Local Resources, 90 J Pub Econ 1519 (2006) (examining county-level
data from 1982 to 1997 estimating that SFJs result in a $48 per capita increase in state education
spending, but a $51 per capita reduction in other intergovernmental transfers from the state).
25 Card and Payne, 83 J Pub Econ at 50 (cited in note 20) (noting that increases in state
aid cause local governments to reduce taxes, generating a net increase of only fifty to sixty-five
cents).
26 Matthew G. Springer, Keke Liu, and James W. Guthrie, The Impact of School Finance
Litigation on Resource Distribution: A Comparison of Court-Mandated Equity and Adequacy
Reforms?, 17 Educ Econ 421, 436–41 (2009).
27 Christopher Berry, The Impact of School Finance Judgments on State Fiscal Policy, in
West and Peterson, eds, School Money Trials 213, 222–33 (cited in note 13).
28 790 SW2d 186 (Ky 1989).
29 See Koski, 55 Hastings L J at 1084–87 (cited in note 19).
30 See William S. Koski, The Evolving Role of the Courts in School Reform Twenty Years
after Rose, 98 Ky L J 789, 790–91 (2009). See generally, for example, Davis v South Dakota,
2011 WL 3849864 (SD); Connecticut Coalition for Justice in Education Funding, Inc v Rell,
990 A2d 206 (Conn 2008); Roosevelt Elementary School District No 66 v Bishop, 877 P2d 806
(Ariz 1994); Skeen v Minnesota, 505 NW2d 299 (Minn 1993); McDuffy v Secretary of Executive
Office of Education, 615 NE2d 516 (Mass 1993); Abbot v Burke, 575 A2d 359 (NJ 1990).
6 The University of Chicago Law Review [79:1
greater resources to obtain a satisfactory level of education.31
However, evidence so far does not demonstrate a statistically robust
difference in the fiscal effects of equity and adequacy cases.32
Likewise, Caroline Hoxby explores variation in SFJ outcomes by
abandoning the SFJ dummy-variable approach33 in favor of
computing economically meaningful parameters of school-finance
equalization plans on a state-by-state basis.34 She emphasizes that
SFJs can have quite different effects depending on the price and
income effects they impose, whereas the dummy-variable estimates
capture at best the average effects. In the context of the present
Article, Hoxby’s most important point is that the common SFJ
dummy-variable approach masks significant heterogeneity in the
experiences of different SFJs.35
In the remainder of this Article, we look to politics as one
explanation for the heterogeneous effects of SFJs across states. We
emphasize that a court’s ruling of the existing system as
unconstitutional is only the first step toward reform. Some court
orders provide instructions for how the legislature should fix the
system, but most orders let state politicians provide the remedy by
redesigning the finance system themselves.36 State leaders must then
craft a new finance system that garners the approval of the state
legislature and governor. Even after the court ruling, the reform
must pass through the usual lawmaking process of the state.
Therefore, even states with identical court rulings may end up with
very different reforms depending on how the legislature and the
governor respond to them. We investigate the relationship between
political conditions and court impact by estimating models that allow
the effects of an SFJ to vary depending on the partisan composition
of the state government at the time of the ruling. While this
approach captures only one of the many possible ways that politics
31 For a discussion of the different underlying conceptualizations of resource distribution
in equity and adequacy lawsuits, see Anna Williams Shavers, Rethinking the Equity vs.
Adequacy Debate: Implications for Rural School Finance Reform Litigation, 82 Neb L Rev 133,
146–60 (2003).
32 See Springer, Liu, and Guthrie, 17 Educ Econ at 421 (cited in note 26); Berry, School
Finance Judgments at 222–27 (cited in note 27).
33 Earlier studies included in econometric models a simple on-off switch—a dummy
variable—to indicate whether a court had issued an SFJ in a given state. This approach
captured only the average effect of SFJs and did not distinguish between legal, political,
economic, or other variations in the SFJ decisions or the context surrounding those decisions.
See, for example, Murray, Evans, and Schwab, 88 Econ L Rev at 801–04 (cited in note 20).
34 See Caroline M. Hoxby, All School Finance Equalizations Are Not Created Equal,
116 Q J Econ 1189, 1191–97 (2001).
35 Id at 1192.
36 See, for example, Rose, 790 SW2d at 215.
2012] Making Courts Matter 7
can influence the implementation of a court-ordered reform, our
results suggest that partisan differences are important.
II. DATA AND EMPIRICAL STRATEGY
A. Data
Our analysis examines information from Martin West and Paul
Peterson about SFJs and their outcomes for judgments issued
between 1988 and 2005. We exclude funding lawsuits not related to
general education funding37 and cases that are not the first SFJ in a
given state during our period of study. For the list of cases included
in our analysis, see Appendix Table 1.
Annual district-level financial and demographic information
comes from the Common Core of Data (CCD)38 and the US Census
Bureau Elementary-Secondary School System Finance Survey (F-33)39
when CCD data was not available (1988–1992 and 2005). Our
analysis considers only local school districts and parts of local
supervisory unions with at least one hundred students, as identified
37 We exclude fifteen cases from West and Peterson based on the content of the cases.
For a complete listing of all of the significant school-finance judgments from 1971 to 2005, see
West and Peterson, eds, School Money Trials at 345–58 (cited in note 13). These cases are
Kasayulie v Alaska, 1999 WL 34793400 (Alaska Super) (facilities); Crane Elementary School
District v Arizona, No 1 CA-CV 04-0076 (Ariz Ct App 2006) (capital funding); Hull v Albrecht,
950 P2d 1141 (Ariz 1997) and 960 P2d 634 (Ariz 1998) (capital funding); Roosevelt Elementary
School District Number 66 v Bishop, 877 P2d 806 (Ariz 1994) (capital funding); Haley v
Colorado Department of Education, No 02-CV-5149 (CO, Denver County District Court, 2002)
(special education); Idaho Schools for Equal Educational Opportunity v Idaho, 129 P3d 1199
(Idaho 2005) (facilities); School Administrative District No 1 v Commissioner, Department of
Education, 659 A2d 854 (Me 1995) (upholding method for reducing funding); Durant v
Michigan, 566 NW2d 272 (Mich 1997) (special education); Board of Education of the City of
Millville v New Jersey Department of Education, 872 A2d 1052 (NJ 2005) (preschool); Abbott v Burke,
748 A2d 82 (NJ 2000) (preschool); Zuni Public School District v New Mexico, No CV-98-14-II (NM,
McKinley County District Court, May 26, 1999) (facilities); Tennessee Small School Systems v
McWherter, 894 SW2d 734 (Tenn 1995); Tennessee Small School Systems v McWherter, 91
SW3d 232 (Tenn 2002) (teacher salaries); and Neeley v West Orange-Cove Consolidated
Independent School District, 176 SW3d 746 (Tex 2005) (addressing property taxes, not school
funding). We include two cases that settled with an agreement for reform: Mock v Kansas, No
91-CV-1009 (KS, Shawnee County District Court, Oct 14, 1991), available at Philip C. Kissam,
Constitutional Thought and Public Schools: An Essay on Mock v. State of Kansas, 31
Washburn L J 474, 489–504 (1992); and Minneapolis Branch of the NAACP v Minnesota, No
95-14800 (Minn, Hennepin County Dist Ct 2000), consolidated with Xiong v Minnesota, No 98-
2816 (Minn, Hennepin County Dist Ct 2000).
38 National Center for Education Statistics, Common Core of Data, online at
http://nces.ed.gov/ccd/bat/ (visited Oct 24, 2011) (providing demographic and fiscal information
about public elementary and secondary schools).
39 US Census Bureau, Public School Finance Data, online at http://www.census.gov/govs
/school/ (visited Oct 24, 2011) (providing finance data, including revenues, expenditures, debt,
and assets, for elementary and secondary public school systems).
8 The University of Chicago Law Review [79:1
by the CCD. To address apparent data errors,40 we follow the
approach of Murray, Evans, and Schwab,41 and therefore exclude
districts with total spending per pupil above the ninety-ninth
percentile for each given state-year.42 Additional district
demographic information, such as the proportion of the population
ages five to seventeen and the proportion of school-aged children
living in poverty, comes from the US Census Bureau’s “Model-
Based Small Area Income and Poverty Estimates” for most years.43
Because district poverty information is not available for every year,
the poverty estimates from the closest available survey year are used.
For example, the district poverty estimates for 1996, 1997, and 1998
all use the data from 1997. State-level political control and
demographic information, including population and per capita
income, come from the US Census Bureau’s “Statistical Abstract of
the United States”44 and the National Governors Association
website.45
Local-school-district spending in the United States is financed
by three sources of funding: local revenues raised by the district,
primarily through property taxes; state aid; and federal aid.
Together, state and local funding constitute over 90 percent of school
funding.46 On average, state and local funding each account for
roughly half of this total, although there is substantial variation
across states in the allocation of funding responsibilities between
state and local governments. Most state spending on education
consists of transfers to local districts rather than direct delivery of
40 For example, one district that had 237 students reported over $500,000 in per-pupil
total funding.
41 Murray, Evans, and Schwab, 88 Am Econ Rev at 795 (cited in note 20).
42 We also exclude two districts with implausibly low funding per pupil, as well as Hawaii
and Washington, DC, because each has only one school district.
43 US Census Bureau, Model-Based Small Area Income and Poverty Estimates, online at
http://www.census.gov/did/www/saipe/ (visited Oct 24, 2011). Additionally, for 1989 and 2005,
the district demographic information comes from National Center for Education Statistics,
School District Demographics System, online at http://nces.ed.gov/surveys/sdds/index.aspx
(visited Oct 24, 2011).
44 US Census Bureau, Statistical Abstract of the United States: 2007 253 table 400, online
at http://www.census.gov/prod/2006pubs/07statab/election.pdf (visited Feb 15, 2011); US
Census Bureau, Statistical Abstract of the United States: 1990 259 table 434, online at
http://www2.census.gov/prod2/statcomp/documents/1990-01.pdf (visited Feb 15, 2011).
45 National Governors Association, Past Governors Bios, online at http://www.nga.org/cms
/home/governors/past-governors-bios.html (visited Oct 24, 2011).
46 National Center for Education Statistics, Revenues and Expenditures for Public
Elementary and Secondary Education: School Year 2008–09 (Fiscal Year 2009) 5 (2011), online
at http://nces.ed.gov/pubs2011/2011329.pdf (visited Oct 24, 2011).
2012] Making Courts Matter 9
educational services.47 The federal government has never provided a
significant portion of funding for public education but rather limits
its role to funding specific programs, such as the Free and Reduced
Price Lunch Program, Title I funding for schools that serve students
living in poverty, and various programs supporting special education.
B. Empirical Strategy
In the analyses that follow, we estimate the impact of SFJs on
state, local, and total funding for school districts. While SFJs require
a policy response from the state government, and therefore are
expected to have a direct impact on state funding, they may also
have an indirect effect on local funding—that is, the revenue raised
by local school districts. Indeed, one concern over the efficacy of
SFJs is that local districts may reduce their own taxes in response to
increases in state aid, thereby undermining state-level efforts to
increase school spending. For example, David Card and Abigail
Payne find that a dollar of increased state aid to a local school
district results in only fifty to sixty-five cents in increased spending.48
Therefore, to provide a more comprehensive picture of the effect of
SFJs, we estimate models of both state and local funding. In
addition, models of total funding capture the net effect of SFJs.49
We use a differences-in-differences strategy to identify the effect
of court-ordered education reform on school district funding.
Specifically, we estimate a series of regression models of the basic
form:
= + + ∙ +
∙ + ∙ +
where represents per-pupil funding (in constant 2007 dollars) for
district located in state at time , and and are district- and
year-fixed effects, respectively. With the inclusion of district-fixed
effects, identification in our models comes from within-district
changes in funding following an SFJ. The fixed effects account for
all, observable or unobservable, time-invariant attributes of the
47 A notable exception is Hawaii, which has the nation’s only entirely state-run school
system. See Hawai’i State Department of Education, Introduction, Organization, online at
http://doe.k12.hi.us/about/intro_org.htm (visited Oct 24, 2011).
48 Card and Payne, 83 J Pub Econ 49, 50 (cited in note 20). But see Baicker and Gordon,
90 J Pub Econ at 1532 (cited in note 24).
49 Because federal aid also contributes to local district spending, total funding is not
simply the sum of state and local funding. However, because we do not expect federal funding
to be influenced by a state SFJ—an expectation that is borne out in the data, see Table 2—we
do not estimate additional models for this variable. Nevertheless, estimates of federal funding
are implicit in the difference between results for total funding and the sum of the results for
state and local funding.
10 The University of Chicago Law Review [79:1
district, including the attributes of the state in which the district is
located. Meanwhile, the year-fixed effects purge the estimates of
national trends that affect all districts, such as economic conditions
or changes in federal education policy. Our main variable of interest
is , which is a dummy variable equal to 1 if the state’s schoolfinance
system has been ruled unconstitutional in a court challenge.
Specifically, the dummy variable is set to 1 in the year after the
decision and each subsequent year.50 In most specifications, we also
interact with state partisan control or district poverty, as will be
explained in the analyses that follow. is a vector of time-varying
district characteristics and is a vector of time-varying state
characteristics, both of which will be specified below. Finally, are
the errors, clustered by state, which account for arbitrary correlation
in the residuals among districts within the same state and serial
correlation of repeated observations from the same district.
We include time-varying state-level characteristics that are
expected to influence state funding of education. We control for the
fraction of the state’s population over sixty-five to account for the
possibility that the elderly oppose increases in school spending.51 We
also control for the fraction of the population that is of school age,
which captures aggregate demand for educational services. At the
same time, increasing per-pupil spending is proportionally more
expensive where children constitute a larger fraction of the
population. We control for per capita income in the state, reflecting
“Wagner’s Law” that demand for government services increases with
income52—that is, the elasticity of government spending with respect
to individual income is greater than 1. We also include the square of
state income to allow for the possibility that the demand for
educational spending is not unbounded. Finally, we include two
dummy variables for years of unified Democratic and unified
Republican control, respectively, of the upper chamber, the lower
chamber, and the governorship, which capture partisan differences in
state education funding. The omitted category is divided
government.
We also control for time-varying characteristics of school districts
that are expected to influence either their receipt of state funding or
their propensity to raise their own local revenue. We include the
50 We allow for a one-year lag for the SFJ to take effect because we assume that any
changes in policy made as a result of the decision will be reflected in the next year’s budget at
the earliest.
51 James M. Poterba, Demographic Structure and the Political Economy of Public
Education, 16 J Pol Analysis & Mgmt 48, 49 (1997).
52 Adolf Wagner, Three Extracts on Public Finance, in Richard A. Musgrave and Alan T.
Peacock, eds, Classics in the Theory of Public Finance 1, 7 (Macmillan 1958).
2012] Making Courts Matter 11
number of students in the district to allow for the possibility that
large districts operate differently from small districts, perhaps due to
economies of scale.53 We control for the percentage of the student
population living in poverty, which we expect to make the district
more likely to receive state funding and less able to raise local
funding. The fraction of students with an individualized education
plan (IEP) is an indicator of special needs, which we expect to result
in higher spending. We also include the proportion of the student
population that is African American and the proportion that is
Hispanic. Although we have no reason to believe that these two
variables directly cause changes in education funding, we have access
to relatively few district covariates and include these due to the
possibility that they are correlated with other relevant factors in the
district, such as property values or population growth.54
Descriptive statistics for all the variables used in our analysis are
presented in Table 1.
III. BASELINE RESULTS
We begin by estimating a set of baseline models that assess the
overall impact of SFJs on education funding. Models 1 through 3 of
Table 2 present the within-district estimates for total, state, and local
funding. The point estimates indicate that state funding (Model 2)
increases following an SFJ by about $775 per pupil, while local
funding (Model 3) declines by roughly $400. Total funding (Model 1)
increases by a little over $300.55 This pattern is broadly consistent
with the results shown by Berry, who finds, using state aggregate
data and a longer time series going back to 1971, that SFJs lead to
increasing centralization of education funding to the state level.56
However, none of the SFJ effects estimated in Models 1 through 3
achieve statistical significance. In other words, although on average
SFJs led to increases in total and state funding and reductions in
local funding, there is substantial variation across states and districts
in the impact of SFJs.
53 See Murray, Evans, and Schwab, 88 Am Econ Rev at 803 (cited in note 20). With the
dependent variable measured in per-pupil terms, a negative coefficient on the number of
students indicates economies of scale.
54 For a more general analysis of the role of district racial composition in school-finance
litigation, see generally James E. Ryan, The Influence of Race in School Finance Reform,
98 Mich L Rev 432 (1999).
55 The change in total funding is not exactly equal to the change in state funding plus the
change in local funding because federal funding, which we do not estimate directly, may also
change after an SFJ. As is implicit in this table and those that follow, changes in federal
funding after an SFJ are a negligible part of the total changes in funding.
56 See Berry, School Finance Judgments at 214, 227–31 (cited in note 27).
12 The University of Chicago Law Review [79:1
Models 4 through 6 isolate one important source of variation in
the impact of SFJs by estimating an interaction between the SFJ
dummy and an indicator for district poverty. Because most lawsuits
were aimed, in one way or another, at increasing funding for poor
districts specifically, there is no reason that an SFJ must affect
funding for the average district. Our estimated interaction term
allows for the impact of an SFJ to be different for poor districts
relative to other districts. Specifically, we use an indicator for
whether a district is in the highest quartile, defined relative to its
state and year, in the proportion of students living in poverty.57 All of
the estimated interactions between district poverty and SFJ status
are highly significant statistically, indicating that SFJs led to changes
in funding disproportionately for poor districts. The total effect of
SFJs on the funding of poor districts is the sum of the SFJ main
effect plus the SFJ times poor interaction, which is reported at the
bottom of the table along with the corresponding standard errors.
Taking this sum, one can see that poor districts received $1,260 more
in state funding, $600 less in local funding, and $640 more in total
funding as a result of SFJs, although the estimates for local revenue
are very imprecise. The net increase of $640 in total funding for poor
districts is fairly modest substantively, considering that total perpupil
funding in our sample is a little over $14,000 on average. Put
differently, $640 represents about one-eighth of a standard deviation
in total per-pupil funding. The results for nonpoor districts follow the
same pattern—a shift from local to state revenue resulting in a
modest increase in total funding—but are even smaller in magnitude
and insignificant statistically.
The main effect of the district-poverty dummy variable—which
can be interpreted as the effect of being in the highest poverty
quartile in the absence of an SFJ—is negative and significant for total
revenue (Model 4). Interestingly, the negative differential in total
funding is due to poor districts receiving less state aid (Model 5),
rather than to differences in locally raised revenue (Model 6). Again,
according to the SFJ-poverty interaction, SFJs lead to significant
increases in state funding, as well as total funding, resulting in poor
districts receiving more total revenue after an SFJ, compared both to
nonpoor districts with an SFJ and to poor districts without an SFJ.
57 Continuous measures of poverty yield comparable results, as will be discussed below.
We begin with the dichotomous variable because we have theoretical reasons to expect that
the relationship between the impact of an SFJ and district-level poverty will not be continuous.
That is, we expect an SFJ to result in increases in funding for poor districts, though not
necessarily reductions in funding for other districts.
2012] Making Courts Matter 13
Overall, the results shown in Models 4 through 6 are consistent
with results from prior studies finding that SFJs result in modest
reductions in the inequality of education funding within a state.58 At
the same time, while these prior studies utilize state-aggregate
measures of inequality, we are the first to produce estimates from
district-level data, demonstrating that SFJs generate increases in
state aid to high-poverty districts specifically.
Among the control variables, the point estimates are generally
in the expected direction, although only a few achieve statistical
significance. Districts spend less per student when enrollment
increases, suggesting the existence of economies of scale. Districts
spend more when the proportion of students with an IEP increases.
Given that IEPs generally apply to students with special needs, this
result is unsurprising. Finally, districts receive more funding in the
years when Republicans control the state government.59 This result is
somewhat surprising in light of the popular perception that
Republicans are the party of fiscal conservatism, although it is
possible that increased education funding comes at the expense of
reductions in other expenditures. We will have more to say about
partisan differences in the next Section.
IV. PARTISAN DIFFERENCES
The results presented in the previous Section recapitulate the
major findings from prior studies—namely, that SFJs lead to
transfers of fiscal responsibilities from local to state governments,
increases in the overall level of education funding, and reductions in
funding inequality. In this Section, we investigate whether the impact
of an SFJ varies depending on the partisan control of the state
government at the time of the decision. We focus on patterns of
partisan control at the time of the decision because the sitting
government is the one obliged to craft a policy response to the
court’s order.60 While later governments are also constrained by the
court’s prior rulings, they are not charged with bringing the schoolfunding
system into compliance in the first place. Ultimately,
whether the government in place at the time of the decision leaves
58 See, for example, Berry, School Finance Judgments at 233 (cited in note 27); Card and
Payne, 83 J Pub Econ at 66 (cited in note 20). But see Murray, Evans, and Schwab, 88 Am
Econ Rev at 806 (cited in note 20). High poverty does not necessarily imply low spending, so
the results in Table 2 do not speak directly to the issue of spending inequality.
59 See Table 2.
60 Of course, there are instances in which the sitting legislature and governor fail to agree
on a response to the court’s order before their term is up, thereby leaving the problem for their
successors to tackle. We discuss these issues below.
14 The University of Chicago Law Review [79:1
an enduring mark on the system is an empirical question, and later
we will discuss the robustness of our results as compared to
alternative ways of modeling partisan influence.
We begin by creating three dummy variables representing,
respectively, unified Democratic control of the state legislature and
governorship at the time of the court decision, unified Republican
control, and divided government. For example, the first dummy
variable will equal 1 for years after an SFJ if Democrats controlled
both houses of the legislature and the governorship in the year of the
decision, 0 otherwise. This approach allows us to test whether the
within-district difference in funding before versus after an SFJ is
different depending on the partisan composition of the state
government at the time of the decision. The omitted category is
again no SFJ. Thus, identification in these models comes both from
within-district variation (to estimate the change in spending after an
SFJ) and from between-state variation (to estimate the differential
effects of the three configurations of partisan control). We continue
to include the full set of state and district control variables used in
Table 2, although to conserve space we do not report their results.
Appendix Table 1 lists the cases used in our analysis and the
configuration of partisan control of the state government at the time
of the decision. From the outset, we caution that there are only three
SFJs issued during periods of unified Republican government.61
There are seven SFJs handed down to unified Democratic
governments,62 and eleven SFJs delivered to divided governments.63
While we have nearly 200,000 district-level observations in our data
set, our ability to test for partisan differences in outcomes is limited
by the number of cases decided under each regime.
Models 1 through 3 of Table 3 indicate that SFJs handed down
to Democratic state governments result in the greatest changes in
education funding. Specifically, when Democrats are in control at the
time of an SFJ, state revenue subsequently increases by $1,750 per
pupil, while local revenue declines by $630, resulting in an increase in
total funding of $1,150. Meanwhile, SFJs implemented under divided
government result in roughly offsetting increases in state funding and
decreases in local funding, with negligible average effects on total
funding. Finally, when a Republican-controlled government
responds to an SFJ, the result is modest increases in both state and
61 These SFJs were issued for New Hampshire, Ohio, and Wyoming.
62 These SFJs were handed down to Alabama, Kentucky, Maryland, Missouri, Tennessee,
Vermont, and West Virginia.
63 These SFJs were delivered to Connecticut, Idaho, Kansas, Massachusetts, Minnesota,
Montana, North Carolina, New Jersey, New York, South Carolina, and Texas.
2012] Making Courts Matter 15
local funding, summing to a total increase in funding of $365 per
student. Most of the coefficients are imprecisely estimated, however.
Only the effect of Democratic reform on total funding is significant
at the conventional 5 percent level, while the Republican effect on
total funding is significant at 10 percent.
The hypothesis tests implicit in Table 3 indicate whether each
type of partisan reform results in a difference in spending compared
with not having had an SFJ. It is also important to test whether the
different configurations of partisan control have different effects
from each other. These hypothesis tests are reported in Table 3A.
The evidence shows that Democratic-controlled reforms do not have
the same total effect as a reform implemented under divided
government. However, all of the other tests fail to reject the
hypothesis that reforms issued under different partisan
configurations have equal effects.
The imprecision of the estimated effects in Models 1 through 3
again suggests that there is substantial heterogeneity across districts
in the effects of SFJs, even under the same regime of partisan
control. In Models 4 through 6, we allow the effects of SFJs to differ
for poor and nonpoor districts within each regime type. That is, we
have a quadruple-difference model in which we estimate the
difference in the change in spending after an SFJ between poor and
nonpoor districts across three different regimes of partisan control.
Using this strategy, it becomes apparent that the reforms
implemented by Democratic state governments have significantly
different effects from those implemented by divided or Republican
governments.
First, note that the Democratic-reform-times-poor-districtinteraction
term is substantively small and statistically insignificant
for total revenue, indicating that poor districts do not fare differently
than other districts under a Democratically controlled SFJ. The main
effect of Democratic reform remains positive and significant. In
other words, all districts—poor and nonpoor alike—see increases in
funding following an SFJ implemented by a Democratic state
government. According to Models 5 and 6, poor districts do experience
a greater shift from local to state funding than nonpoor districts
under a Democratic SFJ, but the decrement to local revenue
essentially offsets the increment to state revenue, hence the null
result for total revenue.
By contrast, SFJs implemented by divided or Republican state
governments generate significant increases in funding for poor
districts, relative to other districts. That is, the main effects for split
and Republican SFJs are insignificant, while the interactions with the
16 The University of Chicago Law Review [79:1
poor-district indicator are positive and significant. Under both types
of regimes, poor districts see reductions in local funding that are
more than offset by increases in state funding, resulting in increases
in total funding. Although the point estimate for the SFJ times poordistrict
interaction is larger for Republican governments ($622) than
for divided governments ($498), we cannot reject the hypothesis that
they are equal, as shown in Table 3A.
In sum, an SFJ under each type of partisan regime results in a
change to the school-funding system after the ruling, and
Democratically implemented SFJs have significantly different effects
from SFJs implemented under Republican or divided state
governments. Specifically, Democratic SFJs result in increases in
total funding for both poor and nonpoor districts, whereas a
Republican or divided government SFJ results in increases in
funding for poor districts disproportionately. All three types of SFJs
lead to a shift in funding responsibility from local to state
governments, though to differing degrees.
Thus far in our analysis, we have lumped together all nonpoor
districts within a state into a single category. To produce a more
comprehensive picture of the partisan effects of SFJs, we next extend
our analysis to include interactions between SFJ status, partisan
regime type at the time of the decision, and indicators for four
quartiles of the within-state distribution of district-level poverty. In
other words, we divide each state’s districts into four quartiles based
on the proportion of students in poverty and then allow each of the
three different types of SFJs—Democratic, Republican, and divided
state government—to have differential effects for each quartile. This
allows us to assess, in a nonparametric fashion, the effect of an SFJ
across the spectrum of district poverty. As before, the inclusion of
district- and year-fixed effects means that our estimates of the effect
of an SFJ come from within-district changes in spending after the
decision. We simply compare these within-district changes across
different combinations of partisan control and district poverty.
The results of this analysis are presented in Table 4. The omitted
category of district poverty is the second-poorest quartile, so the
coefficients for the remaining quartiles should be interpreted as
differences with respect to this omitted category. However, these
complex interactions are more easily presented graphically, which we
do in Figure 1. Each bar in the graph represents the total effect of an
SFJ—that is, the within-district change in spending after the
decision—for each category of partisan control and district poverty,
computed from the coefficients shown in Table 4. Meanwhile,
2012] Making Courts Matter 17
Table 4A presents associated tests for differences in the effects of
SFJs across various partisan and poverty groups.
The more detailed results depicted in Figure 1 indicate that SFJs
crafted by Democratic governments result in total funding increases
across the board. Districts in every poverty quartile see a shift from
local to state funding after an SFJ, with a net increase in funding
ranging from roughly $1,000 to $1,300 per pupil. While a few of the
differences between quartiles are significant statistically, they are
substantively small relative to the overall level of the funding
increases. Indeed, if anything, the results indicate that the most
affluent districts fare better than the poorest districts, in terms of
total funding, after a Democratic SFJ, although this difference is not
statistically significant.
SFJ reforms implemented under divided government produce
decidedly different results. There is a shift from local to state
funding, but the change in state funding is monotonically increasing
district poverty. The poorest quartile of districts receives significantly
greater increases in state funding than any other quartile. Moreover,
the poorest quartile is the only one to see positive net changes in
total funding. Overall, divided government SFJs appear to represent
a more or less straightforward redistribution of funding toward the
poorest districts. The net effects on state education funding appear
to be neutral, as further evidenced by Model 1 of Table 3, showing
little change in total education funding after an SFJ under divided
government. That said, the net increase in total funding for poor
districts is substantively small, amounting to roughly $350 per
student, or only 7 percent of a standard deviation in total funding.
Republican-controlled SFJs present yet a third distinct pattern
of funding reforms. Under Republican reforms, there is a shift from
local to state funding only for the poorest districts. Districts in the
most affluent quartile face cuts in state funding, but they are able to
more than compensate for these reductions by increasing their own
local revenue. In other words, Republican SFJs involve increasing
centralization of funding for the poorest districts and increasing
decentralization of funding for the richest districts. The middle two
quartiles are essentially unaffected. On net, both the poorest and the
richest districts see increases in total funding, the former courtesy of
state aid and the latter financed from their own tax base. Indeed, the
richest districts in Republican states are the only quartile under any
partisan regime to experience an increase in local funding following
an SFJ.
Which partisan arrangement leads to the best results for poor
districts after an SFJ? Clearly, SFJs implemented by Democrats
18 The University of Chicago Law Review [79:1
produce the largest net increases in funding. However, by delivering
roughly equivalent funding increases to districts at all income levels,
Democratic SFJs will do little to reduce the overall level of funding
inequality in their states. On the other hand, SFJs implemented by
divided or Republican governments deliver concentrated benefits to
poor districts, thereby reducing, albeit moderately, the funding
differential between poor and more affluent districts. However,
especially under divided government SFJs, the actual flow of new
dollars into poor districts is more meager than when Democrats are
in control. Assessing which sort of reform produces the greatest
benefits would require a richer theory and empirical analysis of the
connection between student outcomes and the level and inequality of
school funding.
We do not offer a substantive political explanation for the
differential partisan responses to state SFJs. One simple conjecture
would be that political actors use the court rulings as a pretense to
enact policy changes that they were otherwise predisposed to make
anyway. Yet, it is not clear why such a pretense would be necessary
in a state dominated by one party. For example, when Democrats
control the legislature and governorship, they presumably will set the
policy they want absent court intervention, so it is not obvious why
an SFJ should elicit a greater response from Democrats than from
Republicans. It is possible that supermajoritarian voting requirements
within the legislature prevent the dominant party from attaining its
ideal policies even under unified government, and that court rulings
help surmount such hurdles. However, a comprehensive account of
court influence over state policymaking is beyond the scope of this
paper and so we leave the theoretical underpinnings of our results as
fodder for future research.
V. ROBUSTNESS
In additional analyses not reported, we have explored the
robustness of our results to various changes in model specification
and variable definitions. First, we examined sensitivity to our
treatment of partisan control. In the models reported above, we
coded each SFJ according to the partisan alignment in the year of the
decision. As an alternative, we also reran the models using
contemporaneous measures of partisan control—that is, interacting
the SFJ dummy with indicators of partisan control in the current year
for all years after the decision. The signs of the coefficients were
unchanged, but they became much less precisely estimated when
using the contemporaneous measures of partisanship. This leads us
to conclude that state party control at the time of the decision
2012] Making Courts Matter 19
captures an important determinant of the political response to the
court decision.
As noted in our discussion of the school-finance litigation
movement, many states have had more than one court ruling on
education finance.64 The rulings that determine our coding of the SFJ
dummy variable are listed in Appendix Table 1. However, some
states had rulings in place prior to the beginning of our study period,
and some states had additional rulings that came later. We have
explored various sensitivity analyses as to the coding of the SFJ
dummy. For example, we recoded the dummy variable to reflect
cases decided before the beginning of our study period and reran the
models. We also tried including indicator variables for the existence
of reform decisions that came before or after the cases used in our
analysis. Our results did not change importantly, although we did
observe that the first reform case tends to have the largest effect on
fiscal outcomes.
A lingering concern with our results may be that the filing of
lawsuits is endogenously related to political control of the state
government. First, we note that because nearly all states—forty-five
of fifty—were subject to at least one education-finance lawsuit, the
central issue is not whether a state would face a suit but when. One
possibility is that education-reform advocates may time the filing of
their lawsuits to take advantage of what they view as particularly
favorable political conditions. Another possibility is that advocates
might resort to litigation only when the legislative and political
process fails to provide reform. Therefore, endogenous timing by
litigants might potentially bias our results in either direction.
We believe that the possibility of endogenous lawsuit timing visà-
vis political regime is not a major concern for three reasons: the
amount of time between lawsuit filing and the court decision is
unpredictable, political regimes often change between the lawsuit
filing and decision, and lawsuits do not appear to be precipitated by
changes in political regime. Among the twenty-three cases that
contribute to our estimate of SFJ reform, the length of time from the
initial filing through the final appellate court decision ranged from
less than a year to nine years. On average, the process took four
years. Due to the length of time the suits take and the variability of
the speed of the adjudication process, advocates could not effectively
time their lawsuits to specific political regimes. In almost half of the
cases—eleven out of twenty-three—the political regime changed
between the time of filing and time of decision. When political
64 See note 16 and accompanying text.
20 The University of Chicago Law Review [79:1
control changes more rapidly than lawsuits can be adjudicated,
reform advocates cannot time their lawsuits to be decided under a
specific regime. Further, SFJ lawsuits do not appear to be motivated
by a change in regime preceding case filing. On average, the political
regime in the state was stable for six years prior to the filing of a
case. In only three cases did the regime change in the year of the
lawsuit filing, and for each of those three cases, the political regime
changed again before the lawsuit was decided. Thus, we do not see
evidence that endogenous timing of lawsuits is likely to be an
important source of bias for our results.
CONCLUSION
The state experience with school-finance litigation provides
evidence that courts have played a significant role in at least this one
realm of social reform. We find statistically significant differences in
education funding following court judgments overturning a state’s
school-financing system. In particular, Democratic, Republican, and
divided state governments produce significantly different education
spending patterns following an SFJ, once we account for the relative
poverty of the school districts. Further, SFJ reforms overall had a
significantly different impact on low-spending school districts than
on high-spending school districts. The effects of a court decision are
contingent on the pattern of partisan control in place at the time.
Thus, our analysis challenges the idea that courts are ineffectual in
initiating policy change while providing some of the first quantitative
evidence that implementation is contingent on interactions between
courts and other political institutions.
2012] Making Courts Matter 21
FIGURE 1. EFFECT OF SFJ ON SCHOOL FUNDING BY PARTY
CONTROL AND POVERTY QUARTILE
Notes: This is a graphical depiction of the results shown in Table 4. Bars represent the sum of the
coefficients for the partisan SFJ main effect plus its interaction with district poverty. The first bar in each
section indicates the highest poverty quartile, while the fourth bar indicates the lowest poverty quartile.